Torsten Ehlers Frank Packer

torsten.ehlers@bis.org frank.packer@bis.org

Green bond finance and certification1

Financing of investments through green bonds has grown rapidly in recent years. But definitions
of what makes a bond “green” vary. Various certification mechanisms have evolved to allow more
granularity as well as continuity in assessment. Green bonds have been priced at issuance at a
premium on average relative to conventional bonds, but their performance in the secondary
market over time has been similar. A relatively large share of green bonds are in sectors subject
to environmentally related credit risks. More consistent green bond standards across jurisdictions
could help to further develop the market.
JEL classification: G24, O16, Q50.

Green bonds are fixed income securities which finance investments with
environmental or climate-related benefits. Green bonds are an integral component
of “green finance” more generally, which aims to “internalize environmental
externalities and adjust risk perceptions” for the sake of increasing environmentally
friendly investments (G20 GFSG (2016)). Economic theory teaches that a first-best
solution for closing the gap between the private and social costs of pollution would
be a mix of lump sum taxes and subsidies, with regulations to impose implicit prices
following closely behind. Green finance can also help to alleviate these externalities,
through market-based means. It acts by increasing the flow of funds to
environmentally beneficial projects, essentially reducing their costs, as well as by
heightening awareness of the financial risks related to environmental change.2
How can investors be sure that the proceeds of green bonds are invested in an
environmentally friendly way, and not merely “green-washed” to give the appearance
thereof? While there is no single global definition of what precisely constitutes an
“environmentally beneficial” use of proceeds, different standards have gained
acceptance among market participants. Various organisations have started to provide
green label certifications that indicate adherence to particular definitions of green,
including “shades” of green. In so doing, they align the incentives of those who want
to invest in these bonds, and make it easier for asset managers to satisfy those
preferences. Green bonds could also conceivably serve as a hedge against

1
The views expressed in this article are those of the authors and do not necessarily reflect those of the
BIS. We thank Claudio Borio, Stijn Claessens, Benjamin Cohen, Dietrich Domanski, Kumar
Jegarasasingam, Luiz Pereira da Silva and Hyun Song Shin for their valuable comments, and Agne
Subelyte for providing excellent research assistance. We also thank the Climate Bonds Initiative for
providing us with their data.
2
See the proceedings of the conference co-organised by the Official Monetary and Financial
Institutions Forum (OMFIF), the BIS and the World Bank Group in Frankfurt on 13 July 2017; see also
Weidmann (2017) and Pereira da Silva (2017).

BIS Quarterly Review, September 2017 89

environmentally related financial risks,3 though in this case additional information is
needed about the sensitivity of various bonds to such risks, beyond just the quality
of “greenness” itself .
In this feature we provide an overview of the state of the green bond market. We
show that, a decade into the development of the market, there still are numerous
labels for green bonds. We highlight how the various certification mechanisms for
greenness have been evolving so as to allow more granularity as well as continuity in
assessment. Consistent with other research, we document that green bonds at
issuance have been priced at a premium on average relative to conventional bonds,
while their performance in the secondary market has been similar to other bonds if
currency risks are hedged. Finally, we document that green bonds are exposed to
environmentally related financial risks to a relatively high degree.
The rest of this feature proceeds as follows. After providing a definition of green
bonds, we briefly review the growth and composition of labelled green bond issuance
from two different sources. The third section examines and classifies the various green
labels provided by the private sector to certify green bond issuance. The fourth
section focuses on the market pricing, financial performance and risks of green bonds.
The conclusion summarises the policy implications.

The market for green bonds

The market for bonds with a green label4 has grown rapidly in recent years. It started
with the European Investment Bank’s “climate awareness bond” issued in 2007, which
is widely seen as the first bond with a green label. A key catalyst for subsequent
market development was the introduction in January 2014 by the International
Capital Market Association (ICMA) of the Green Bond Principles, which are the basis
for many of the existing green labels (ICMA (2014)). Since then, the market for labelled
green bonds has expanded dramatically: in 2016, aggregate issuance surpassed the
$100 billion mark, and the first half of 2017 has already seen a total issuance of around
$60 billion (Graph 1). The market for green bonds is nevertheless still very small
compared with the wider global bond market, with a share of less than 1.6% of global
debt issuance in 2016.5
The composition of green bond issuance has evolved considerably over time
(Graph 1, left-hand panel). Through 2013, issuance predominantly came from
supranationals (ie international organisations such as the European Investment Bank).
Issuers from advanced economies in Europe and the United States dominated in 2014
and 2015. Since 2016, issuers from emerging market economies (EMEs), particularly

3
Transition risks, ie the financial impact of changes in environmental regulation, are typically seen as
the most significant risk bond investors face. Carney (2015) notes that “while a given physical
manifestation of climate change – a flood or storm – may not directly affect a corporate bond’s value,
policy action to promote the transition towards a low-carbon economy could spark a fundamental
reassessment”.
4
The universe of bonds for which proceeds are used for environmentally beneficial projects, but which
do not carry a green label, is still likely to be quite large (Climate Bonds Initiative (2015)). But as the
green label becomes more commonly used, it becomes more and more likely that issuers will seek
this label to distinguish themselves.
5
Based on an estimate of total global debt capital market issuance of $6.69 trillion in 2016
(see www.dealogic.com/insights/key-trends-shaped-markets-2016/).

90 BIS Quarterly Review, September 2017

Rapid growth in green bond issuance most recently in EMEs
Climate Bonds Initiative and Bloomberg lists of labelled green bonds, in billions of US dollars Graph 1

Green bond issuance by country of incorporation Green bond issuance since 20071
Years USD bn USD bn

8 90 150

6 60 100

4 30 50

2 0 0
1
07 08 09 10 11 12 13 14 15 16 17 Local currency Foreign currency Supranational
Average Rhs: Supranationals China US dollar Other AE currencies5
maturity United States Other EMEs
3
Euro EM currencies6
(lhs) Germany OFCs
4

France
2
Other AEs
1
Up to mid-June 2017. 2 AT, AU, BE, CA, CH, DK, ES, EE, FI, GB, IT, JP, LU, LV, NL, NO, SE and SI. 3 AE, AR, BR, CL, CO, HK, IN, IR, KR, MX,
PE, PH, PL, SG, TR, TW and ZA. 4 Offshore financial centres (OFCs): BM, CR, KY, MO, MU and VG. 5 AUD, CAD, CHF, GBP, JPY, NOK, NZD
and SEK. 6 BRL, CNY, COP, HUF, IDR, INR, MAD, MXN, MYR, PEN, PHP, PLN, RUB, SGD, TRY, TWD and ZAR.

Sources: Bloomberg; Climate Bonds Initiative; authors’ calculations.

corporates in China, have provided a large share of global issuance. The maturity of
green bonds is typically medium-term, averaging between seven and eight years in
2014–16, but lengthened in the first half of 2017.

Issuance tends to cover a wide range of currency denominations. While

supranationals have tended to issue in euros and dollars, some EME entities have
made use of local currencies (Graph 1, right-hand panel).
Issuers of green bonds tend to be highly rated, with only a small fraction rated
below investment grade (Graph 2, left-hand panel). More bonds with earmarked
claims on the cash flows of an individual project (ie project bonds) could help diversify
issuers and credit risks (Standard & Poor’s (2017a), Caldecott (2012); Ehlers et
al (2014)). Project bonds naturally entail higher credit risks, given the high initial
uncertainties of new investment projects (Ehlers (2014)). These could cater to
investors looking for higher yields.
The above numbers aggregate figures from two widely used and broad-based
lists of green bonds: those of the Climate Bonds Initiative (CBI) and Bloomberg. The
CBI is an international non-profit organisation, funded by grants from non-profit and
government sources as well as revenue from public sector contracts. Bloomberg is a
widely used financial data and service provider. Both lists selectively include bonds
with different types of green labels.
The two lists, and hence the corresponding volumes of issuance, do not coincide.
The aggregate issuance of labelled green bonds from 2007 through the second
quarter of 2017 in the CBI list equals $234 billion (comprising 1,092 individual issues).
By contrast, Bloomberg counted $216 billion in green bond issuance (comprising
779 individual issues). Only $169 billion (624 individual issues) worth of bonds are

BIS Quarterly Review, September 2017 91

Distribution of green bond ratings and amounts issued1
In per cent Graph 2

Share of credit ratings2 Share of total issuance volume

100 100

80 80

60 60

40 40

20 20

0 0
2013 2014 2015 2016 2017 2013 2014 2015 2016 2017
AAA A+ to BBB– CBI only Bloomberg only Both
AA+ to AA– Below BBB–
1
For 2017, issuance data up to mid-June. 2
Average rating of Standard & Poor’s, Moody’s and Fitch; expressed in Standard & Poor’s
categories.

Sources: Bloomberg; Climate Bonds Initiative (CBI); authors’ calculations.

present on both lists. The Bloomberg list of green bonds was less extensive prior to
2014, but since then the overlap has been greater (Graph 2, right-hand panel).

Forms of green bond certification

For investment in green bonds to take off, it is important for both asset managers
and their principals to be able to identify the bonds that actually have environmental
or climate-related benefits. Asset managers may have the resources to make an
informed judgment on their own. Indeed, global initiatives such as the Financial
Stability Board’s (FSB) Task Force on Climate-Related Financial Disclosures aim to
make better environmental information readily available (FSB TCFD (2016)). Still,
external certification allows asset managers to show beneficiaries that they are indeed
investing in green bonds when requested to do so, and may be more cost-effective.
A variety of forms of green bond certification have emerged, which all aim at ensuring
that the use of funds and subsequent revenue is tied to green investment (Table 1).
The ICMA Green Bond Principles are so-called “voluntary process guidelines” that
outline general criteria that most certification schemes follow. They were put together
by major private financial institutions under the aegis of the ICMA (ICMA (2015)). The
Principles provide prospective issuers with guidance on the key components of green
bond issuance, namely: (i) the use of proceeds for environmentally sustainable
activities; (ii) a process for determining project eligibility; (iii) management of the
proceeds in a transparent fashion that can be tracked and verified; and (iv) annual
reporting on the use of proceeds.
Though this section reviews international certification mechanisms available to
any issuer, many jurisdictions have developed their own national taxonomies of what
constitutes eligibility as a green bond. Most notably, China’s Green Bond Finance
Committee has issued a Green Bond Endorsed Project Catalogue (People’s Bank of

92 BIS Quarterly Review, September 2017

Characteristics of different green bond identification and certification schemes Table 1

CBI Climate Green bond CICERO Second Moody’s Green Standard &
Bonds indices1 Opinions Bond Poor’s Green
Certification Assessments Evaluations
Use of funds must be tied to green Yes Yes Yes Yes Yes
investment
Eligibility criteria differ by sector Yes Yes Yes
Ex post monitoring/assessment Yes
Granular assessments of greenness Yes Yes Yes
Quantitative weights for specific factors Yes Yes
1
Bank of America Merrill Lynch, Barclays MSCI, Standard & Poor’s and Solactive.

China (2015)). For large domestic markets, this is a sensible option, but to the extent
that international harmonisation is an issue, domestic guidelines run the risk of
limiting the value of any particular green certification scheme to the domestic
investor base. The ongoing initiative to improve the consistency of definitions and
methodologies for determining the eligibility of green projects across the
jurisdictions of China and the European Union represents perhaps the most
significant effort to address this issue (European Investment Bank (2017)).

Climate Bond Certification

As discussed above, the CBI maintains a list and database of green bonds issued since
2009. The bonds in its database have green labels, but inclusion in the database does
not constitute an opinion by CBI as to the correctness of the label.
The CBI also provides standards and a certification procedure. While the Green
Bond Principles are very general, the CBI’s Climate Bonds Standard establishes sector-
specific eligibility criteria to judge an asset’s low carbon value and suitability for
issuance as a green bond. Assets that meet the CBI standard are then eligible for
Climate Bond Certification, after an approved external verification that the bond
meets environmental standards and that the issuer has the proper controls and
processes in place.
A limitation of the CBI standard is that it does not necessarily mandate
monitoring and verification on an ongoing basis (Table 1, first column). It is highly
useful for investors to have an entity regularly renew its certification, particularly if
they intend to maintain the investment over a multi-year horizon.

Green bond indices

Green bond indices identify specific bonds as green via a stated methodology, and
allow investors to invest in a portfolio of green bonds to diversify risks (Table 1,
second column). To this extent, the green bond index providers also effectively act as
institutions of certification. At present, global green bond indices are compiled by
Bank of America Merrill Lynch, Barclays MSCI, Standard & Poor’s and Solactive.6 Each

6
In addition, there are several internationally listed green bond indices focusing on specific
jurisdictions – in particular China. For instance, the Shanghai Green Bond Index Series (developed by

BIS Quarterly Review, September 2017 93

has its own methodology for choosing the components of the index. While
advertising consistency with the Green Bond Principles, each index also specifies
additional factors such as size and liquidity, as well as the specific industry sectors for
which the proceeds are used.
Index providers can arguably serve an ongoing monitoring function, since they
can discard entities from an index as well as include them. However, since many
inclusion criteria for green bond indices are much less concrete than those for
conventional bonds (such as minimum levels of market liquidity and credit ratings), it
remains to be seen whether the index providers can monitor such environmental
criteria on a continuous basis.

External reviews

Though it is not a part of the four main Green Bond Principles, it was recommended
in the 2015 edition of the Principles that green bond issuers “use external assurance
to confirm alignment with the key features of Green Bonds”. This can include second
opinions and verifications. From 2016, the Principles referred to “external reviews”
rather than “external assurance”, while the list of recommended external reviews was
expanded to include those provided by rating agencies (ICMA (2016)).
A limitation of the CBI standard and inclusion in the green bond indices is their
binary nature: a bond is either green or not. More granular assessments could contain
valuable information for investors, such as the degree of environmental benefits, or
whether environmental benefits are likely to persist. Indeed, the following providers
of green certification make more granular assessments.
CICERO is a climate research institute based in Oslo (Table 1, third column) and
the leading provider of second opinions. It evaluates the issuer’s framework for both
project selection and investment (CICERO (2016)). CICERO provides three different
degrees of positive assessment (“shades of green”), reflecting the bond’s adherence
to a long-term vision for a low-carbon, “environmentally resilient” society. However,
CICERO reviews the green bond framework at the time of issuance. Ex post changes
in the framework or environmental impact are not monitored, unless the issuer
specifically requests it.7
Moody’s Green Bond Assessments. The first public methodology for the
assessment of green bonds by a ratings agency was published by Moody’s Investors
Service in March 2016 (Table 2, second column). Green Bond Assessments (GBAs) are
intended to “assess the relative likelihood that bond proceeds will be invested to
support environmentally friendly projects”, in line with the Green Bond Principles
(Table 2, first column). As with their credit rating products, Moody’s employs
numerous quantifiable factors in determining the GBAs, with the explicit aim of

the Shanghai Stock Exchange in collaboration with China Securities Index Co), or the CUFE-CNI Green
Bond Index Series (developed by the Shenzhen Securities Information Co together with the
International Institute of Green Finance) are indices based on green bonds issued in China which are
also listed in Europe on the Luxembourg stock exchange.
7
More regular monitoring is now implicitly recommended by the Green Bond Principles though not
necessarily delegated to third parties. In 2016, the Principles were updated to state that issuers should
“... keep readily available up to date information on the use of proceeds to be renewed annually until
full allocation and as necessary thereafter in the event of new developments”.

94 BIS Quarterly Review, September 2017

Comparison of Green Bond Principles and rating agencies’ green certifications Table 2

Green Bond Principles
Use of proceeds
Use should be described and
present clear environmentally
sustainable benefits.
Process for project evaluation
and selection
Decision-making process should
be outlined; in particular, how
projects fit into green
categories, eligibility criteria and
environmental sustainability
objectives.
Management of proceeds
Net proceeds should be tracked
by formal internal process.
Reporting
Issuers should provide annual
list of projects to which
proceeds are allocated.
1
The methodology of Standard & Poor’s Green Evaluations differs between instruments that finance mitigation projects and those that
finance adaptation projects. The comparison of the table focuses only on the evaluations of instruments that finance mitigation projects,
or those that “bring environmental benefits and target areas such as natural resources depletion, loss of biodiversity, pollution control,
and climate change”.
Moody’s Green Bond Assessments Standard & Poor’s Green Evaluations1
Use of proceeds Mitigation
Assessment depends on percentage of Assesses the environmental impact of
proceeds allocated to eligible project financing proceeds over the life of the
categories. Weighted 40% as a factor. assets. Weighted 60% as a factor.
Organisation Governance
Sub-factors: effectiveness of Considers whether well defined
environmental governance and procedures in place for:
organisation structure; rigorous review - Selecting projects eligible to be
and decision-making process; qualified financed
personnel and/or reliance on third parties;
- Appraising and managing
explicit criteria for investment selection;
environmental impact
external evaluations for decision-making.
Weighted 15% as a factor. - Complying with environmental
regulations
Weighted 19% as a factor.
Management of proceeds Governance
Sub-factors: segregation and tracking of Considers whether well defined
proceeds on accounting basis; tracking of procedures in place for preventing
the application of proceeds by proceeds of the bond from being used
environmental category and project type; for other purposes than the intended
reconciliation of planned investments green financings.
against allocations; eligibility rules for Weighted 6% as a factor.
investment cash balances; external or
independent internal audit.
Weighted 15% as a factor.
Disclosure on use of proceeds Transparency
Sub-factors: description of green projects; - Use of proceeds reporting
adequacy of funding to complete - Impact reporting and disclosure
projects; quantitative descriptions of
- External verification of impact data
targeted environmental results; methods
and criteria for calculating performance Weighted 15% as a factor.
against targets; reliance on external
assurances. Weighted 10% as a factor.
Ongoing reporting and disclosure
Sub-factors: reporting and disclosure
post-issuance; ongoing annual reporting;
granular detail on nature of investment
and environmental impact; quantitative
assessment of impacts to date;
comparison of assessments of impacts
with projections at time of issuance.
Weighted 20% as a factor.

Sources: International Capital Market Association (2015); Moody’s Investors Service (2016a); Standard & Poor’s (2017b).

BIS Quarterly Review, September 2017 95

increasing their transparency and replicability. Regular review is anticipated, similar
to the way regular credit ratings are refreshed over a multi-year horizon.8
Standard & Poor’s Green Evaluations. Standard & Poor’s introduced Green
Evaluations in 2017 (Table 2, third column). The focus of these ratings is broader than
that of GBAs, as they include a technical environmental impact assessment
component, along with governance and transparency components. A score between
zero and 100 is intended to evaluate the relative ranking of the overall expected
lifetime environmental impact relative to maintaining the status quo – after
discounting for qualities of the governance and transparency of the bond’s use of
proceeds.
Notwithstanding its broader focus, the S&P methodology can also be viewed as
mostly in line with the Green Bond Principles, while providing transparency with
regard to the quantitative importance of various factors, and considerable granularity
in the final assessment. That said, each score is a point-in-time evaluation, and is
removed from the S&P website after at most 18 months. In this sense, ex post
assessment is not provided unless requested as part of a separate evaluation.

The pricing of green bonds, financial performance and risks

For the green bond market to channel a significant amount of funds into
environmentally friendly projects, green bonds should also fulfil the needs of both
issuers and investors. Looking at the same issuer, the risk characteristics of a green
bond are essentially identical to those of a conventional bond: while the proceeds
from the issuance of a green bond are earmarked for environmentally friendly
projects, green bonds are serviced from the cash flows of the entire operations of the
issuer – not just the green project.
These characteristics have implications for the pricing of green bonds and their
attractiveness for investors. A premium at issuance over comparable bonds without
a green label would indicate that a significant number of investors value the label,
enough to give issuers an extra incentive to issue bonds that have it. At the same
time, these investors will still be interested in an acceptable financial performance of
green bonds over time. Another consideration is the exposure to credit risks related
to environmental change. That green bonds support environmentally beneficial
projects does not necessarily imply lower exposure to such risks. The rest of this
section discusses these issues further.

The green bond premium

Does the green label influence the price that issuers are willing to pay (in other words,
the yield spread over risk-free rates that they are willing to accept) for a bond? A large
body of literature documents that factors unrelated to credit risks can significantly

8
ESG ratings, which assess a firm’s performance across a weighted average of environmental, social
and governance (ESG) issues, provide many of the same attributes. In June 2016, the Green Bond
Principles statement by ICMA recognised that the “use of proceeds” bond concept might be applied
to themes beyond the environment, such as bonds financing projects with social objectives, or with
a combination of social and environmental objectives. However, the green bond label is still reserved
for investment projects providing clear environmental benefits. Since ESG ratings are not specifically
focused on environmental benefits, they are not considered in this feature.

96 BIS Quarterly Review, September 2017

Credit spreads at issuance of green versus non-green bonds1
In basis points Graph 3

Individual green bond issues by currency denomination Average yield at issuance premia by rating

20 0

0 –10

–20 –20
Average = –18 bp
–40 –30

–60 –40

–80 –50
2014 2015 2016 2017 AAA AA A to BBB
US dollar Euro
1
Relative differences in credit spreads at issuance compared with a non-green bond of the same issuer at the closest possible date. The average
closest date of a non-green bond issue by the same issuer is around seven days before the issue date of the green bond. Sample was restricted
to pari passu bonds of at least two-year maturity at issuance, at least $10 million issuance amount and currency of denomination being either
euros or US dollars. Credit spreads are calculated as the spreads of yields at issuance over the yield curve of US Treasury securities (for US dollar-
denominated bonds) and German Federal Treasury securities (for euro-denominated bonds) of the same maturity at issuance date; if the same
maturity was not available, the available points of the respective yield curve were interpolated.

Sources: Board of Governors of the Federal Reserve System; Deutsche Bundesbank; Bloomberg; authors’ calculations.

influence bond yield spreads, such as specific demand and supply factors (eg Collin-
Dufresne et al (2001), Greenwood and Vayanos (2014)) or liquidity premia
(eg Longstaff (2004), Amihud et al (2006)). If a sizeable population of investors is
willing to pay a premium (accept a lower spread) for green bonds, this should show
up in the pricing of the bonds at issuance.

To analyse the price effect of the green label, we compare the credit spreads at
issuance of a cross-section of 21 green bonds issued between 2014 and 2017 with
the credit spreads at issuance of conventional bonds of the same issuers at the closest
possible issue date (Graph 3). As most issuers of green bonds also regularly issue
conventional bonds, this comparison allows us to control for issuer-specific
idiosyncratic factors, including credit risk. We do not include any project bonds, as
claims on cash flows could possibly be on different projects with different risk
characteristics. We further restrict the sample to pari passu fixed rate bonds, to avoid
the influence that debt seniority or the uncertainty of floating rates could have on the
pricing at issuance. We look for conventional bonds of roughly the same maturity,
and restrict our matched pairs to US dollar- and euro-denominated green bonds, as
spreads over local benchmark rates tend not to be as stable for bonds issued in EME
currencies.
Our results indicate that green bond issuers on average have borrowed at lower
spreads than they have through conventional bonds. This confirms the results from
other recent studies (eg Zerbib (2017), Barclays (2015)). The mean difference in spread
in our sample is around 18 basis points.9 Overall, this is consistent with a high

9
An 18 basis point lower credit spread would be significant relative to the potential costs of a green
label or rating. The certification fee for the green label of the Climate Bonds Initiative is a flat 0.1 basis
points of the issue value (though the CBI also requires the external engagement of a party that verifies

BIS Quarterly Review, September 2017 97

demand for green bonds (CBI (2017), OECD (2016)) relative to supply – in other words,
enough investors have a preference for holding green bonds to influence the issue
price. When we segment the sample by rating category, it appears that the green
yield difference is greater for riskier borrowers (Graph 3, right-hand panel). Recall that
we compare credit spreads at issuance of pari passu bonds from the same issuer, so
this result does not reflect differences in credit risk (or other factors) across issuers
within the same rating category. Again, this finding is consistent with previous work
(Zerbib (2017)).
At the same time, we also document considerable variation across the individual
green bond issues in our sample (Graph 3, left-hand panel). The standard deviation
of the premium is 27 basis points. Not all issuers were able to take advantage of a
yield discount at issuance: five out of the 21 green bond issues priced at spreads
above the matched conventional bonds. We also could not document yield premia
for higher levels of greenness as determined by the more granular assessments of
the major rating agencies.10

Financial performance over time

Returns to investors from green bonds will reflect their performance over time, in
particular when investors do not intend to simply hold them until maturity. The
average premium at issuance will not necessarily translate into a noticeable
underperformance in secondary market trading. Investors in the secondary market
may well price in a different premium from primary market investors.11 Many
investors will also be interested in the realised volatility of green bonds as a metric of
financial performance.
Green bond indices are a good starting point to analyse the secondary market
performance of green bonds from an investor perspective. Green bond indices
contain a diversified broader portfolio of bonds12 and thereby provide a good means
of comparison with the performance of other bond indices that are suitable for a wide
range of investors. We analyse so-called hedged returns, which measure returns in
US dollars that can be achieved by hedging the currency exposures of the underlying
index. As green bond indices differ notably from other global bond indices in their
currency composition, currency movements alone can have an outsize impact on
relative returns.

procedures and reports). As for the green assessments of the major rating agencies, even if they were
to be as expensive as a normal credit rating (3–5 basis points of the issue volume (White (2002))), the
costs would be far less than 18 basis points.
10
These new assessments have so far tended to be at one end of the spectrum: for instance, the 26
bond issues with Moody’s GBAs as of end-July 2017 were all at the highest level (GB1). The 10
available bond issues with green rating evaluations by Standard & Poor’s rank between 67 and 92
(out of 100), but the only four of these bonds for which we could find yields are from the same issuer
and therefore have the same score. This does not allow us to make meaningful comparisons of the
pricing of bonds across different green bond ratings. Likewise, because of a paucity of yield data, we
cannot compare the impact of the rating agency assessments with those of the other green labels.
11
Buy and hold investors will have decided ex ante that the benefits of holding the individual issue to
maturity equal or outweigh the premium.
12
Some analysts, however, have argued that green bond indices are less diversified than other much
broader bond indices (Fitch Ratings (2017)). Nevertheless, the green bond indices we consider
represent an accurate overall picture of the financial characteristics of the green bond market at its
current stage of development.

98 BIS Quarterly Review, September 2017

Green bond indices: return characteristics
Annualised monthly total returns, July 2014–June 2017; in per cent Table 3

Hedged returns1 Unhedged returns2

Cumulative Mean Std Sharpe Cumulative Mean Std Sharpe
over 36 dev ratio over 36 dev ratio
months (1) (2) = (1)/(2) months (1) (2) = (1)/(2)

Green Standard & Poor’s . . . . –4.18 –1.48 18.44 –0.08

bond BofA Merrill Lynch 9.17 3.06 8.64 0.35 –6.45 –2.15 18.92 –0.11
indices
Solactive . . . . –4.52 –1.51 18.66 –0.08
Barclays MSCI 10.23 3.41 9.61 0.35 –3.43 –1.14 17.23 –0.07
Global Broad-based 11.26 3.75 11.64 0.32 3.46 1.15 14.97 0.08
bond
AA average rating 10.29 3.43 10.21 0.34 0.97 0.32 14.16 0.02
indices
A average rating 11.02 3.67 10.99 0.33 3.12 1.04 14.08 0.07
1
Total returns for indices hedged against currency risks, which are more comparable across indices that differ in their currency
composition. They represent a close estimation of the return that can be achieved by hedging the currency exposures of the underlying
index by selling foreign currency forwards at one-month forward rates. Hedged returns are not exact measures of the true returns, since
they do not include transaction costs for hedging contracts. Those can be considered small, in particular, since euro exposures present
the largest foreign currency component (see also Graph 1). Bid-ask spreads on EUR/USD forwards, which are an indicator of transaction
costs, are usually between 1 and 1.5 bp. 2 Total returns from holding a simple position in the respective index. Green bond indices
differ substantially in their currency composition from the global bond indices. Hence, currency movements will significantly affect return
differentials between green bond and global bond indices.

Sources: Bloomberg; Barclays; authors’ calculations.

Overall, the performance of hedged green bond indices has been similar to that
of global bond indices of comparable credit rating composition13 (Table 3, left-hand
columns). Even though the available green bond indices by Bank of America Merrill
Lynch, Barclays MSCI, Standard & Poor’s and Solactive diverge slightly in their
composition (CBI (2015)), their return characteristics hardly differ. The ratio of average
monthly hedged returns to their standard deviation (Sharpe ratio), which is a standard
measure for risk-adjusted performance, was in some cases even slightly higher for
green bond indices than for global bond indices, though that difference was not
statistically significant.
Total returns in US dollars on unhedged green bond indices, however, have
exhibited higher volatility than those of broad-based bond indices (Table 3, right-
hand columns). As the currency composition of the green bond indices is very diverse,
and much less tilted towards the US dollar than is the case for global indices,
movements in the currency of denomination of some green bonds in the indices
increase the measured return volatility. This points to the importance of the
availability of currency hedges for investors in green bonds.14

13
The universe of labelled green bonds from Bloomberg and CBI has an average rating of slightly below
AA. The average maturity and duration of the global bond indices and the different green bond
indices are quite similar, but sectoral and currency compositions differ.
14
Importantly, some researchers (eg Andersson et al (2016)) have shown that it is possible to construct
a portfolio of green instruments that has an identical performance to that of a broader market index.
This would provide a product for a wide range of green investor that yields a financial performance
equal to that of a diversified market portfolio, but with environmental benefits.

BIS Quarterly Review, September 2017 99

The green label and exposure to environmental risks
Carney (2015) describes the potentially severe impact of climate change on the
economy as a “tragedy of the horizon”: investors and other actors do not sufficiently
take the risks into account as they materialise only over a long horizon. Indeed, a
number of academic studies have documented a tendency for investors to
inadequately price environmental risks (Hong et al (2016) and references therein).
This is despite already strong evidence of severe financial impacts of both physical
risks, due to climate-related events such as droughts and floods, and transition risks,
such as the risk of a material change in environmental regulations (Caldecott et
al (2014)). For instance, rating agencies now plan to take account of the financial risks
related to the transition to stricter carbon emission rules implied by the Paris
agreement when they analyse credit risks of bond issuers from polluting sectors
(eg Moody’s Investors Service (2016b)).
One question is whether green bonds can provide an instrument for investors to
hedge against these environmentally related financial risks. If these risks materialise,
bonds from issuers in polluting sectors may be subject to significant revaluations. To
the extent that issuers of green bonds are better shielded against large revaluations,
they could serve as an efficient risk management instrument.
But this need not be the case. Green bonds generally comprise investor claims
on an entity’s overall operations. A large and diversified energy company may invest
a considerable amount in green projects, but other parts of its business (for instance
coal power plants) may expose it to environmentally related credit risks, such as
changes in carbon regulations. There may also be green bonds whose income stream
is vulnerable to climate change quite apart from transition risk (for example, wind
farms subject to flood risk). In the case of almost all green bonds, the exposure to
environmentally related credit risks is a function of the entire company’s business.
Only a few green bonds are project bonds, where claims are on the cash flows of the
financed green project itself.

Environmental credit risk composition of all rated bonds versus green bonds only1
In per cent Graph 4

Total rated debt (2015) Green bonds

Share of total issuance since 20072
0.7
2.2 3.7

10.3 10.1
77.6
86.7
8.6

Immediate, Elevated Risk Emerging, Elevated Risk Emerging, Moderate Risk Low Risk
1
Aggregate issuance of bonds by issuers from sectors which belong to the risk categories shown, as defined by Moody’s. See Moody’s
Investors Service (2015). In cases where industrial classification is ambiguous, we use equal weights to distribute the issuance volume across
relevant sectors. 2 Through mid-June 2017.

Sources: Bloomberg; Climate Bonds Initiative; Moody’s; authors’ calculations.

100 BIS Quarterly Review, September 2017

 Indeed, the evidence suggests that green bonds are more, not less, exposed to
environmentally related credit risks. Moody’s provides a classification of credit
exposures to environmental risks at the sectoral level (Moody’s Investors
Service (2015) and Annex Table A1).15 Within the universe of corporate debt rated by
Moody’s, 13.2% is issued by institutions in industries with moderate or greater
exposure to environmental credit risk, and around 2.9% in industries classified as
either immediate or emerging elevated risk (Graph 4, left-hand panel). By contrast,
when we examine the industry composition of green bonds alone, we see that 22.4%
of green bonds are issued in sectors with moderate or greater exposure to
environmental credit risk, and nearly 14% in industries classified as elevated risk
(right-hand panel). Thus, the percentage of green bonds in high-risk sectors exceeds
that for overall rated debt by a factor of four.

Conclusion
Green bond principles and standards are an important step towards promoting green
finance. Since the introduction of the Green Bond Principles by the ICMA in January
2014, the issuance of labelled green bonds has increased rapidly, with a growing
number of issuers from the private sector and EMEs. Several green bond indices have
also been introduced, allowing a broader group of investors to take a diversified
position in green bonds. The evidence suggests that investors place value on the green
label at issuance, even though the post-issuance financial performance of green bonds
is comparable with that of conventional bonds. However, for this still relatively small
market to grow more, several further developments need to take place.
First, the various existing definitions and labels for green bonds pose a challenge
for investors, who may benefit from more consistent standards. The ongoing work to
improve the consistency of standards in China and the European Union are promising
in this regard. At the same time, more ongoing monitoring by “second opinion”
providers, rating agencies or other forms of continuous third-party verification may
be needed. Even if asset managers utilise the green label simply to signal to ultimate
investors their fulfilment of green mandates, the information value of those labels can
depreciate over time as technology evolves or policies of the issuer change.
A second informational aspect that is not covered by current green certification
schemes is the environmentally related financial risks of green bonds. While the
management of environmental risks extends far beyond green bonds, it is important
to avoid the misperception that green bonds are insulated from such risks. In fact,
among all rated bonds, those with a green label are more likely to be in sectors that
are exposed to such risks. Green bond standards could be enhanced to highlight the
degree of financial risks stemming from environmental factors so as to further
encourage investors to manage these risks effectively.

15
Credit risks from environmental exposures are defined as the risks to a borrower’s ability to repay
caused by physical climate events or changes in environmental regulations.

BIS Quarterly Review, September 2017 101

References

Amihud, Y, H Mendelson and L Pedersen (2006): “Liquidity and asset prices”,

Foundations and Trends in Finance, vol 1, no 4, pp 269–364.
Andersson, M, P Bolton and F Samama (2016): “Hedging climate risk”, Financial
Analysts Journal, vol 72, no 3, pp 13–32.
Barclays (2015): “The cost of being green”, US Credit Focus, September.
Caldecott, B (2012): “Green infrastructure bonds: accessing the scale of low cost
capital required to tackle climate change”, Climate Change Capital.
Caldecott, B, J Tilbury and C Carey (2014): “Stranded assets and scenarios”, Smith
School of Enterprise and the Environment, Discussion Paper, January.
Carney, M (2015): “Breaking the tragedy of the horizon – climate change and financial
stability”, speech given at Lloyd’s of London, 29 September.
CICERO (2016): Framework for CICERO’s ‘Second Opinions’ on green bond investments,
April.
Climate Bonds Initiative (2015): Bonds and climate change – the state of the market in
2015, July.
——— (2017): Green bond pricing in the primary market: Q4 2016 snapshot, March.
Collin-Dufresne, P, R Goldstein and J Spencer Martin (2001): “The determinants of
credit spread changes”, Journal of Finance, vol LVI, no 6, December, pp 2177–207.
Ehlers, T (2014): “Understanding the challenges for infrastructure financing”, BIS
Working Papers, no 454, August.
Ehlers, T, F Packer and E Remolona (2014): “Infrastructure and corporate bond markets
in Asia”, in Financial flows and infrastructure financing, Reserve Bank of Australia,
conference volume, July.
European Investment Bank (2017): “New People’s Bank of China and EIB initiative to
strengthen green finance”, press release, 22 March.
Fitch Ratings (2017): “Green bond funds face diversification challenge”, special report,
August.
FSB Task Force on Climate-Related Financial Disclosures (2017): Final report:
recommendations of the task force on climate-related financial disclosures, June.
G20 Green Finance Study Group (2016): Green finance synthesis report 2016,
September.
Greenwood, R and D Vayanos (2014): “Bond supply and excess bond returns”, Review
of Financial Studies, vol 27, no 3, pp 663–713.
Hong, H, F Li and J Xu (2016): “Climate risks and market efficiency”, NBER Working
Papers, no 22890.
International Capital Market Association (2014): Green Bond Principles, January.
——— (2015): The Green Bond Principles 2015, March.
——— (2016): The Green Bond Principles 2016, June.
Longstaff, F (2004): “The flight-to-liquidity premium in US treasury bond prices”, The
Journal of Business, vol 77, no 3, pp 511–26.

102 BIS Quarterly Review, September 2017

Moody’s Investors Service (2015): “Environmental risks – heat map shows wide
variations in credit impact across sectors”, September.
——— (2016a): Green Bonds Assessment (GBA), March.
——— (2016b): “Moody’s to analyse carbon transition risk based on emissions
reduction scenario consistent with Paris agreement”, June.
Organisation for Economic Co-operation and Development (2016): “Mobilising bond
markets for a low-carbon transition”, Green finance and investment, April.
People’s Bank of China (2015): “Announcement No. 39,” December.
Pereira da Silva, L (2017): “Green finance: can it help combat climate change?”,
remarks at the Global Public Investor Symposium on “Green bond issuance and other
forms of low-carbon finance”, Frankfurt am Main, 13 July, organised by the BIS,
OMFIF, the Deutsche Bundesbank and the World Bank Group.
Standard & Poor’s (2017a): “We won't solve for green finance unless we solve for
infrastructure”, S&P Global Market Intelligence, April.
——— (2017b): “Green evaluation analytical approach”, S&P Global Market
Intelligence, April.
Weidmann, J (2017): “Green bond issuance and other forms of low-carbon finance”,
speech at the Global Public Investor Symposium on “Green bond issuance and other
forms of low-carbon finance”, Frankfurt am Main, 13 July, organised by the BIS,
OMFIF, the Deutsche Bundesbank and the World Bank Group.
White, L (2002): “The credit rating industry: an industrial organisation analysis”, in
R Levich, G Majnoni and C Reinhart (eds), Ratings, rating agencies and the global
financial system.
Zerbib, O (2017): “The green bond premium”, July.

BIS Quarterly Review, September 2017 103

Annex

Moody’s sectoral environmental risk classification

Credit exposure to environmental risks Table A1

Environmental risk Description Relevant industry examples

classification
Immediate, Sectors are already experiencing material credit Unregulated utilities and power
Elevated Risk implications as a result of environmental risk. Rating companies.
changes have either already been occurring for a
substantial number of issuers or ratings changes are
likely within the next three years.
Emerging, Sectors overall have clear exposure to environmental Automobile manufacturers, power
Elevated Risk risks that, in aggregate, could be material to credit generation projects.
quality over the medium term (three to five years), but
are less likely in the next three years.
Emerging, Sectors have a clear exposure to environmental risks that Regulated electric and gas utilities
Moderate Risk could be material to credit quality in the medium to long with generation, sovereigns –
term (five or more years) for a substantial number of developing countries, environmental
issuers. However, it is less certain that the identified risks services and waste management,
will develop in a way that is material to ratings for most paper and forest products.
issuers.
Low Risk Sectors in this category have either no sector-wide Supranationals and sovereigns –
exposure to meaningful environmental risks or, if they developed countries, banks and
do, the consequences are not so likely to be material to finance companies, consumer goods,
credit quality. semiconductors and technology
hardware.
Source: Moody’s Investors Service (2015).

104 BIS Quarterly Review, September 2017